A vertically integrated company owns increasing amounts of production of their end product. Apple is vertically integrated, they have products where they make the CPU, the device, the OS, and retail the devices. But ultimately they only sell the end products.

If Apple were to start selling CPUs, you could argue that they had a similar model, but they don't (and won't because their CPUs are a competitive advantage).

What Amazon and SpaceX are doing isn't vertical integration in the classic sense. While Amazon sells AWS, they also have an online retail platform. Their retail platform consumes AWS services. Similarly, SpaceX is both the supplier of launch services and the consumer of launch services.

Amazon's retail operations are profitable, but by being their own best customer, they help Amazon drive profitability and demand for AWS.

What's different is that the entire commercial space launch business isn't very big. I think it's been somewhere in the ballpark of a couple dozen launches a year. SpaceX could conceivably satisfy it just by building one rocket and flying it over and over again. (Granted they don't have upper stage re-usability yet, and the Falcon 9 boosters have a limited lifespan, but they're working to get there with Starship.) So, if they want to have a viable business that involves a large fleet of reusable rockets they need to become their own customer to make sure the rockets don't just sit idle.

Maybe this sort of thing happens all the time with new companies in new markets. The space launch market though has been around for a long time. To me it seems unusual for an industry that's been around for half a century or so to not be big enough to absorb the excess capacity of a newcomer.

Vertically integrating space logistics with space products is pretty cool.